Nvidia’s rise to prominence in high-performance computing is a near-perfect embodiment of the American dream. Three friends, including Taiwanese immigrant Jensen Huang, get fed-up with life at a big computer firm and decide to risk everything to compete in the crowded field of computing hardware. They hatch their plan over a meal at Denny’s in 1993, because Huang worked as a busboy at one of their diners as a youth. Huang himself thinks their pitch is an “unfundable idea” with a “0% chance of success.” After they find funding anyway, the new firm faces numerous setbacks and barely survives. It then roars back, eventually becoming a $3 trillion company now driving the artificial intelligence (AI) revolution and the U.S. stock market’s record-setting gains.
Yet, the Biden administration now plans to punish the company for this stunning success. On September 3, the Department of Justice sent a subpoena to Nvidia, deepening the agency’s antitrust-related probe into the firm’s business practices, and the AI market more generally. This comes on the heels of European officials’ raiding Nvidia’s corporate offices overseas in a separate antitrust investigation.
What is astonishing about this response is that these regulators likely weren’t even aware of the company until recently. Before a few years ago, Nvidia was just the maker of advanced 3D graphics processing units (GPUs) for gamers. Though Nvidia’s stock is riding high today, it lingered under $1 per share for many years.
But then the firm pivoted as the crypto and AI revolutions took off. It realized that it could provide a new breed of high-powered GPUs desperately needed to meet the world’s ever-expanding data-crunching demands. Over the past decade, Nvidia’s turn has helped fuel a $335 billion private-sector-led explosion of AI investment—more than three times China’s financial stake in the industry.
At a time when many are worried about China’s growing algorithmic and computational capabilities, Nvidia’s role in expanding America’s dominance in AI should win it praise and admiration from political leaders. Instead, the Biden White House seems intent on punishing it for getting too big, too quickly.
The Biden antitrust crew does not seem to appreciate the wisdom of Joseph Schumpeter, who, 80 years ago, taught us to reject short-term market snapshots and appreciate how the most important competition comes from unexpected places. It is “competition from the new commodity, the new technology, the new source of supply, the new type of organization,” that really matters, Schumpeter argued, because that sort of innovation “strikes not at the margins of the profits and the outputs of the existing firms but at their foundations and their very lives.”
Schumpeterian competition is a race for the prize of profitability, which drives entrepreneurs like Nvidia’s founders to take bold risks to unseat market leaders with superior innovations. To deny successful firms that prize—what Schumpeter referred to as “the most powerful engine of that progress”—would act as a massive disincentive for current or future companies to take those risks at all. That’s a recipe for technological stagnation and lower economic productivity and growth.
Nvidia is the perfect example of how Schumpeter’s “perennial gales of creative destruction” still buffet the tech economy in unrelenting, unpredictable ways. No wonder, then, that Huang regularly expresses how “worried and concerned” he is about the many other competitors looking to eat his lunch. The unofficial Nvidia motto is, “Our company is thirty days from going out of business.”
Huang is right to be nervous. High-performance computing competition is intense. “The key is that there are a lot of options there,” says Lisa Su, CEO of AMD, one of Nvidia’s leading rivals. “I think we’re going to see a situation where there’s not only one solution. There will be multiple solutions,” she told reporters last December. Su noted that chip design is becoming more sophisticated by the day. Amazon, Apple, Google, Meta, Microsoft, and Tesla are also developing their own AI hardware solutions to power the next generation of their computational and robotic technologies.
While antitrust authorities apparently worry about Nvidia’s “full-stack” computing dominance in AI hardware and software, the reality is that every segment of the AI ecosystem is booming. A recent Forbes survey of the world’s 50 most promising AI innovators included over 1,900 submissions, more than double last year’s count, leading the magazine to conclude, “The artificial intelligence sector has never been more competitive.” FirstMark’s tenth annual map of the AI and machine learning business landscape featured a record 2,011 corporate logos for 2024, including 578 new entrants and many new business segments. The first version in 2012 had just 139 logos and a few market segments.
Contrast Nvidia’s rise with Intel’s recent stumbles. Intel once played the same disruptive innovator role, driven by former CEO Andy Grove’s famous belief that “only the paranoid survive” in a competitive world. Today, by contrast, Intel’s motto might as well be “only the politically connected thrive.” The firm is practically joined at the hip with the Biden Commerce Department and its industrial-policy spending spree following passage of the massive CHIPS and Science Act of 2022.
That bloated law allocated $39 billion in direct grants for semiconductor investments, plus an additional $75 billion in loans and loan guarantees. Intel has been the recipient of much of that largesse and has positioned itself as a sort of semiconductor carrier of last resort for the nation. Yet, Intel currently finds itself “mired in a sales slump worse than anticipated and hemorrhaging cash,” while also apparently considering more drastic moves like cutting back factory expansion plans that government money was supposed to boost.
Some have even suggested that a “CHIPS 2” bill might be needed. But that would just further politicize AI and computing markets, especially as Washington looks to attach various DEI strings to the grants, including labor and childcare-related mandates, plus the same environmental-permitting baggage that holds back so many other infrastructure projects. High-tech industrial policy is inefficient enough as it is; even its supporters admit that adding “woke” mandates to the mix will make it even more costly and ineffective.
Intel and the tech industry should have listened to T. J. Rodgers, former CEO of Cypress Semiconductor. Back in 2000, Rodgers begged his fellow Silicon Valley innovators to avoid “normalizing relations” with big government. “The political scene in Washington is antithetical to the core values that drive our success in the international marketplace and risks converting entrepreneurs into statist businessmen,” he argued. Other tech leaders in those heady days took the same approach. Steve Jobs actively shunned Washington, for example, and Apple only bothered setting up a serious Washington, D.C., policy shop years after he died. Today, every major digital technology leader (with the notable exception of Elon Musk) now sends armies of lobbyists to the Beltway.
Could it be that, by shunning political games, Nvidia has set itself up for a fall? The Chinese Communist Party would love to see that happen, of course, and has invested massively in expanding its own AI capabilities, especially in military applications. China is working to develop its own “national champions” to rival Nvidia in pursuit of the CCP’s stated goal of overtaking America and becoming the world’s AI leader by 2030. Beijing might already be leading us in that race, except for the fact that the CCP has tied the hands of domestic AI innovators with a different set of restraints: on free expression. Like the Internet, AI systems depend on the free flow of information to blossom fully. Beijing cannot tolerate such freedom and has moved to enforce its censorship regime at scale, hobbling its own domestic AI capabilities in the process.
This gives the U.S. a major strategic edge in both the AI industry and in geopolitics. To capitalize on that advantage, however, America must avoid heavy-handed AI regulations and antitrust threats. As the Lexington Institute has noted in recent reports, U.S. national security requires a robust, innovative technology sector as part of a “new arsenal of democracy.” When U.S. companies prosper, they not only boost America’s global competitiveness but also help its more pluralistic values, rather than those of the CCP, shape global information platforms and markets.
Unfortunately, the Biden administration is undercutting those goals when it looks to punish AI market leaders like Nvidia, while also over-regulating AI more generally through executive decrees. In a sprawling, 100-plus-page October 2023 AI executive order, Biden moved to convert the 74-year old Defense Production Act into a regulatory edict to limit algorithmic innovation, even though the law was meant to expand industrial production. This followed the administration’s 2022 AI Bill of Rights, a dour document that insisted algorithmic systems were “deeply harmful” and would “exacerbate discrimination” without extensive preemptive controls from bureaucrats. The entire effort was overseen by “AI czar” Kamala Harris, who might pursue this “algorithmic fairness” agenda even more vigorously if she wins the presidency in November. If so, it would be the greatest gift China could hope for as it looks to catch up in the global AI race.
Photo: Costfoto/Future Publishing via Getty Images